Disentangling rodent behaviors to improve automated behavior recognition

Automated observation and analysis of behavior is important to facilitate progress in many fields of science. Recent developments in deep learning have enabled progress in object detection and tracking, but rodent behavior recognition struggles to exceed 75-80% accuracy for ethologically relevant behaviors. We investigate the main reasons why and distinguish three aspects of behavior dynamics that are difficult to automate. We isolate these aspects in an artificial dataset and reproduce effects with the state-of-the-art behavior recognition models. Having an endless amount of labeled training data with minimal input noise and representative dynamics will enable research to optimize behavior recognition architectures and get closer to human-like recognition performance for behaviors with challenging dynamics.

CatWalk XT gait parameters: a review of reported parameters in pre-clinical studies of multiple central nervous system and peripheral nervous system disease models

Automated gait assessment tests are used in studies of disorders characterized by gait impairment. CatWalk XT is one of the first commercially available automated systems for analyzing the gait of rodents and is currently the most used system in peer-reviewed publications. This automated gait analysis system can generate a large number of gait parameters. However, this creates a new challenge in selecting relevant parameters that describe the changes within a particular disease model. Here, for the first time, we performed a multi-disorder review on published CatWalk XT data. We identify commonly reported CatWalk XT gait parameters derived from 91 peer-reviewed experimental studies in mice, covering six disorders of the central nervous system (CNS) and peripheral nervous system (PNS). The disorders modeled in mice were traumatic brain injury (TBI), stroke, sciatic nerve injury (SNI), spinal cord injury (SCI), Parkinson's disease (PD), and ataxia. Our review consisted of parameter selection, clustering, categorization, statistical evaluation, and data visualization. It suggests that certain gait parameters serve as potential indicators of gait dysfunction across multiple disease models, while others are specific to particular models. The findings also suggest that the more site-specific the injury is, the fewer parameters are reported to characterize its gait abnormalities. This study strives to present a clearly organized picture of gait parameters used in each one of the different mouse models, potentially helping novel CatWalk XT users to apply this information to similar or related mouse models they are working on.

Emerging Role of Translational Digital Biomarkers Within Home Cage Monitoring Technologies in Preclinical Drug Discovery and Development

In drug discovery and development, traditional assessment of human patients and preclinical subjects occurs at limited time points in potentially stressful surroundings (i.e., the clinic or a test arena), which can impact data quality and welfare. However, recent advances in remote digital monitoring technologies enable the assessment of human patients and preclinical subjects across multiple time points in familiar surroundings. The ability to monitor a patient throughout disease progression provides an opportunity for more relevant and efficient diagnosis as well as improved assessment of drug efficacy and safety. In preclinical in vivo animal models, these digital technologies allow for continuous, longitudinal, and non-invasive monitoring in the home environment. This manuscript provides an overview of digital monitoring technologies for use in preclinical studies including their history and evolution, current engagement through use cases, and impact of digital biomarkers (DBs) on drug discovery and the 3Rs. We also discuss barriers to implementation and strategies to overcome them. Finally, we address data consistency and technology standards from the perspective of technology providers, end-users, and subject matter experts. Overall, this review establishes an improved understanding of the value and implementation of digital biomarker (DB) technologies in preclinical research.

Measuring Behavior in the Home Cage: Study Design, Applications, Challenges, and Perspectives

The reproducibility crisis (or replication crisis) in biomedical research is a particularly existential and under-addressed issue in the field of behavioral neuroscience, where, in spite of efforts to standardize testing and assay protocols, several known and unknown sources of confounding environmental factors add to variance. Human interference is a major contributor to variability both within and across laboratories, as well as novelty-induced anxiety. Attempts to reduce human interference and to measure more "natural" behaviors in subjects has led to the development of automated home-cage monitoring systems. These systems enable prolonged and longitudinal recordings, and provide large continuous measures of spontaneous behavior that can be analyzed across multiple time scales. In this review, a diverse team of neuroscientists and product developers share their experiences using such an automated monitoring system that combines Noldus PhenoTyper® home-cages and the video-based tracking software, EthoVision® XT, to extract digital biomarkers of motor, emotional, social and cognitive behavior. After presenting our working definition of a "home-cage", we compare home-cage testing with more conventional out-of-cage tests (e.g., the open field) and outline the various advantages of the former, including opportunities for within-subject analyses and assessments of circadian and ultradian activity. Next, we address technical issues pertaining to the acquisition of behavioral data, such as the fine-tuning of the tracking software and the potential for integration with biotelemetry and optogenetics. Finally, we provide guidance on which behavioral measures to emphasize, how to filter, segment, and analyze behavior, and how to use analysis scripts. We summarize how the PhenoTyper has applications to study neuropharmacology as well as animal models of neurodegenerative and neuropsychiatric illness. Looking forward, we examine current challenges and the impact of new developments. Examples include the automated recognition of specific behaviors, unambiguous tracking of individuals in a social context, the development of more animal-centered measures of behavior and ways of dealing with large datasets. Together, we advocate that by embracing standardized home-cage monitoring platforms like the PhenoTyper, we are poised to directly assess issues pertaining to reproducibility, and more importantly, measure features of rodent behavior under more ethologically relevant scenarios.

How the COVID-19 pandemic highlights the necessity of animal research

Recently, a petition was offered to the European Commission calling for an immediate ban on animal testing. Although a Europe-wide moratorium on the use of animals in science is not yet possible, there has been a push by the non-scientific community and politicians for a rapid transition to animal-free innovations. Although there are benefits for both animal welfare and researchers, advances on alternative methods have not progressed enough to be able to replace animal research in the foreseeable future. This trend has led first and foremost to a substantial increase in the administrative burden and hurdles required to make timely advances in research and treatments for human and animal diseases. The current COVID-19 pandemic clearly highlights how much we actually rely on animal research. COVID-19 affects several organs and systems, and the various animal-free alternatives currently available do not come close to this complexity. In this Essay, we therefore argue that the use of animals is essential for the advancement of human and veterinary health.

Reproducibility via coordinated standardization: a multi-center study in a Shank2 genetic rat model for Autism Spectrum Disorders

Inconsistent findings between laboratories are hampering scientific progress and are of increasing public concern. Differences in laboratory environment is a known factor contributing to poor reproducibility of findings between research sites, and well-controlled multisite efforts are an important next step to identify the relevant factors needed to reduce variation in study outcome between laboratories. Through harmonization of apparatus, test protocol, and aligned and non-aligned environmental variables, the present study shows that behavioral pharmacological responses in Shank2 knockout (KO) rats, a model of synaptic dysfunction relevant to autism spectrum disorders, were highly replicable across three research centers. All three sites reliably observed a hyperactive and repetitive behavioral phenotype in KO rats compared to their wild-type littermates as well as a dose-dependent phenotype attenuation following acute injections of a selective mGluR1 antagonist. These results show that reproducibility in preclinical studies can be obtained and emphasizes the need for high quality and rigorous methodologies in scientific research. Considering the observed external validity, the present study also suggests mGluR1 as potential target for the treatment of autism spectrum disorders.

An Integrated System for the Automated Recording and Analysis of Insect Behavior in T-maze Arrays

Host-plant resistance to insects like thrips and aphids is a complex trait that is difficult to phenotype quickly and reliably. Here, we introduce novel hardware and software to facilitate insect choice assays and automate the acquisition and analysis of movement tracks. The hardware consists of an array of individual T-mazes allowing simultaneous release of up to 90 insect individuals from their individual cage below each T-maze with choice of two leaf disks under a video camera. Insect movement tracks are acquired with computer vision software (EthoVision) and analyzed with EthoAnalysis, a novel software package that allows for automated reporting of highly detailed behavior parameters and statistical analysis. To validate the benefits of the system we contrasted two Arabidopsis accessions that were previously analyzed for differential resistance to western flower thrips. Results of two trials with 40 T-mazes are reported and we show how we arrived at optimized settings for the different filters and statistics. The statistics are reported in terms of frequency, duration, distance and speed of behavior events, both as sum totals and event averages, and both for the total trial period and in time bins of 1 h. Also included are higher level analyses with subcategories like short-medium-long events and slow-medium-fast events. The time bins showed how some behavior elements are more descriptive of differences between the genotypes during the first hours, whereas others are constant or become more relevant at the end of an 8 h recording. The three overarching behavior categories, i.e., choice, movement, and halting, were automatically corrected for the percentage of time thrips were detected and 24 out of 38 statistics of behavior parameters differed by a factor 2-6 between the accessions. The analysis resulted in much larger contrasts in behavior traits than reported previously. Compared to leaf damage assays on whole plants or detached leaves that take a week or more to complete, results were obtained in 8 h, with more detail, fewer individuals and higher significance. The potential value of the new integrated system, named EntoLab, for discovery of genetic traits in plants and insects by high throughput screening of large populations is discussed.

Automated video tracking of thrips behavior to assess host-plant resistance in multiple parallel two-choice setups

BACKGROUND

Piercing-sucking insects cause severe damage in crops. Breeding for host-plant resistance can significantly reduce the yield losses caused by these insects, but host-plant resistance is a complex trait that is difficult to phenotype quickly and reliably. Current phenotyping methods mainly focus on labor-intensive and time-consuming end-point measurements of plant fitness. Characterizing insect behavior as a proxy for host-plant resistance could be a promising time-saving alternative to end-point measurements.

RESULTS

We present a phenotyping platform that allows screening for host-plant resistance against Western flower thrips (WFT, Frankliniella occidentalis (Pergande)) in a parallel two-choice setup using automated video tracking of thrips behavior. The platform was used to establish host-plant preference of WFT with a large plant population of 345 wild Arabidopsis accessions and the method was optimized with two extreme accessions from this population that differed in resistance towards WFT. To this end, the behavior of 88 WFT individuals was simultaneously tracked in 88 parallel two-choice arenas during 8 h. Host-plant preference of WFT was established both by the time thrips spent on either accession and various behavioral parameters related to movement (searching) and non-movement (feeding) events.

CONCLUSION

In comparison to 6-day end-point choice assays with whole plants or detached leaves, the automated video-tracking choice assay developed here delivered similar results, but with higher time- and resource efficiency. This method can therefore be a reliable and effective high throughput phenotyping tool to assess host-plant resistance to thrips in large plant populations.

Automated image-based tracking and its application in ecology

The behavior of individuals determines the strength and outcome of ecological interactions, which drive population, community, and ecosystem organization. Bio-logging, such as telemetry and animal-borne imaging, provides essential individual viewpoints, tracks, and life histories, but requires capture of individuals and is often impractical to scale. Recent developments in automated image-based tracking offers opportunities to remotely quantify and understand individual behavior at scales and resolutions not previously possible, providing an essential supplement to other tracking methodologies in ecology. Automated image-based tracking should continue to advance the field of ecology by enabling better understanding of the linkages between individual and higher-level ecological processes, via high-throughput quantitative analysis of complex ecological patterns and processes across scales, including analysis of environmental drivers.

Overexpression of synphilin-1 promotes clearance of soluble and misfolded alpha-synuclein without restoring the motor phenotype in aged A30P transgenic mice

Lewy bodies and neurites are the pathological hallmark of Parkinson's disease. These structures are composed of fibrillized and ubiquitinated alpha-synuclein suggesting that impaired protein clearance is an important event in aggregate formation. The A30P mutation is known for its fast oligomerization, but slow fibrillization rate. Despite its toxicity to neurons, mechanisms involved in either clearance or conversion of A30P alpha-synuclein from its soluble state into insoluble fibrils and their effects in vivo are poorly understood. Synphilin-1 is present in Lewy bodies, interacting with alpha-synuclein in vivo and in vitro and promotes its sequestration into aggresomes, which are thought to act as cytoprotective agents facilitating protein degradation. We therefore crossed animals overexpressing A30P alpha-synuclein with synphilin-1 transgenic mice to analyze its impact on aggregation, protein clearance and phenotype progression. We observed that co-expression of synphilin-1 mildly delayed the motor phenotype caused by A30P alpha-synuclein. Additionally, the presence of N- and C-terminal truncated alpha-synuclein species and fibrils were strongly reduced in double-transgenic mice when compared with single-transgenic A30P mice. Insolubility of mutant A30P and formation of aggresomes was still detectable in aged double-transgenic mice, paralleled by an increase of ubiquitinated proteins and high autophagic activity. Hence, this study supports the notion that co-expression of synphilin-1 promotes formation of autophagic-susceptible aggresomes and consecutively the degradation of human A30P alpha-synuclein. Notably, although synphilin-1 overexpression significantly reduced formation of fibrils and astrogliosis in aged animals, a similar phenotype is present in single- and double-transgenic mice suggesting additional neurotoxic processes in disease progression.

A 3D analysis of flight behavior of Anopheles gambiae sensu stricto malaria mosquitoes in response to human odor and heat

Female mosquitoes use odor and heat as cues to navigate to a suitable landing site on their blood host. The way these cues affect flight behavior and modulate anemotactic responses, however, is poorly understood. We studied in-flight behavioral responses of females of the nocturnal malaria mosquito Anopheles gambiae sensu stricto to human odor and heat. Flight-path characteristics in a wind tunnel (flow 20 cm/s) were quantified in three dimensions. With wind as the only stimulus (control), short and close to straight upwind flights were recorded. With heat alone, flights were similarly short and direct. The presence of human odor, in contrast, caused prolonged and highly convoluted flight patterns. The combination of odor+heat resulted in longer flights with more landings on the source than to either cue alone. Flight speed was greatest (mean groundspeed 27.2 cm/s) for odor+heat. Odor alone resulted in decreased flight speed when mosquitoes arrived within 30 cm of the source whereas mosquitoes exposed to odor+heat maintained a high flight speed while flying in the odor plume, until they arrived within 15 cm of the source. Human odor evoked an increase in crosswind flights with an additive effect of heat at close range (<15 cm) to the source. This was found for both horizontal and vertical flight components. However, mosquitoes nevertheless made upwind progress when flying in the odor+heat generated plume, suggesting that mosquitoes scan their environment intensively while they progress upwind towards their host. These observations may help to improve the efficacy of trapping systems for malaria mosquitoes by (1) optimizing the site of odor release relative to trap entry and (2) adding a heat source which enhances a landing response.

Motor control in complex regional pain syndrome: a kinematic analysis

This study evaluated movement velocity, frequency, and amplitude, as well as the number of arrests in three different subject groups, by kinematic analysis of repetitive movements during a finger tapping (FT) task. The most affected hands of 80 patients with complex regional pain syndrome (CRPS) were compared with the most affected hands of 60 patients with Parkinson disease (PD) as well as the nondominant hands of 75 healthy control (HC) subjects. Fifteen seconds of FT with thumb and index finger were recorded by a 60-Hz camera, which allowed the whole movement cycle to be evaluated and the above mentioned movement parameters to be calculated. We found that CRPS patients were slower and tapped with more arrests than the two other groups. Moreover, in comparison with the hands of the HC subjects, the unaffected hands of the CRPS patients were also impaired in these domains. Impairment was not related to pain. Dystonic CRPS patients performed less well than CRPS patients without dystonia. In conclusion, this study shows that voluntary motor control in CRPS patients is impaired at both the affected as well as the unaffected side, pointing at involvement of central motor processing circuits.

Reliability analysis of observational data: problems, solutions, and software implementation

In the reliability analysis literature, little attention has been given to the various possible ways of creating a basis for the comparison required to compute observer agreement. One needs this comparison to turn a sequential list of behavioral records into a confusion matrix. It is shown that the way to do this depends on the research question one needs to answer. Four methods for creating a basis for comparison for the computation of observer agreement in observational data are presented. Guidelines are given for computing observer agreement in a way that fits one's goals. Finally, we discuss how these methods have been implemented in The Observer software. The Observer 4.1 supports all the methods that have been discussed. Most of these methods are not present in any other software package.